For years, Artificial Intelligence (AI) methods such as Fuzzy Systems, Genetic Algorithms (GA), Artificial Neural Networks (ANNs), and Knowledge-Based Systems (KBSs) have been employed to make computational devices that can adapt their operation to match changing environment needs. At Essex University we are especially interested in applying such techniques to pervasive computing to create, so-called ambient intelligence environments. In these environments agents are employed both to control the environment and manage the association between agents. A particular problem is that such agents usually require AI implementations with powerful inference mechanisms but posses relatively diminished information storage and processing capability. My research is investigating the feasibility to utilize various types of ANNs (self-organizing, spiking, EFuNN ANNs etc) for the construction of such embedded agents.